This repository contains the implementation of the code described in the paper: https://www.mdpi.com/2075-1680/13/2/127.
Please note that the specifications in this repository differ from those in the paper. The new experiments were conducted on the following hardware:
- CPU: Intel i7-13700F
- Base Frequency: 2.1 GHz
- Turbo Frequency: 5.2 GHz
- Cores: 16 (8 performance cores + 8 efficient cores)
- L1 Cache: 1.4 MB
- L2 Cache: 24.0 MB
- L3 Cache: 30.0 MB
- GPU: NVIDIA GeForce RTX 4070TI
- Memory: 12 GB GDDR6X, 21008 MHz
- CUDA Cores: 7680
- RAM: 32 GB DDR5, 4800 MHz, Dual Channel (16+16 GB)
- CUDA Version: 12.5
In the paper, the following table of reference is provided for the networks generated:
| ID | No. of Instances | No. of Nodes | No. of Paths | Erdős–Rényi Probability |
|---|---|---|---|---|
| 1 | 10000 | 1000 | 500 | 0.5 |
| 2 | 10000 | 1000 | 500 | 0.7 |
| 3 | 10000 | 1000 | 500 | 0.9 |
| 4 | 1000 | 2000 | 1000 | 0.1 |
| 5 | 1000 | 2000 | 1000 | 0.15 |
| 6 | 1000 | 2000 | 1000 | 0.6 |
| 7 | 1000 | 2000 | 1000 | 0.6 |
| 8 | 100 | 5000 | 2500 | 0.1 |
| 9 | 100 | 5000 | 2500 | 0.2 |
| 10 | 100 | 5000 | 2500 | 0.3 |
| 11 | 5 | 10000 | 5000 | 0.15 |
| 12 | 5 | 10000 | 5000 | 0.3 |
| 13 | 5 | 10000 | 5000 | 0.5 |
| 14 | 3 | 15000 | 7500 | 0.15 |
| 15 | 2 | 20000 | 7500 | 0.15 |
| 16 | 1 | 25000 | 8000 | 0.15 |
Table 1: Network generator parameters used for experiments.
Due to complications (too many arcs generated) during path generation, this project uses the number of arcs generated as a reference. Therefore, the number of paths will be much lower, but this does not mean that only the given number of paths are present; it merely ensures the minimum.
* indicates missing data in my notes. The C++ code used in the paper was not versioned. I will attempt to recover it from my old computer, or, if necessary, rewrite it.
If you notice discrepancies in the configurations compared to the paper's parameters, it is because I am following the specifications in Table 2, which focus on the actual size of the networks.
| No. of Nodes | Erdős–Rényi Probability | No. of Arcs |
|---|---|---|
| 1000 | 0.5 | 502,949 |
| 1000 | 0.7 | 702,955 |
| 1000 | 0.9 | 900,158 |
| 2000 | 0.1 | 406,696 |
| 2000 | 0.15 | 608,077 |
| 2000 | 0.6 | 2,413,862 |
| 2000 | 0.6 | 2,526,737 |
| 5000 | 0.1 | 5,041,275 |
| 5000 | 0.2 | 7,556,983 |
| 5000 | 0.3 | 15,124,614 |
| 10000 | 0.15 | 15,124,614 |
| 10000 | 0.3 | 20,161,784 |
| 10000 | 0.5 | 50,356,783 |
| 15000 | 0.15 | 34,007,566 |
| 20000 | 0.15 | 60,441,434 |
| 25000 | 0.15 | * |
| 30000 | 0.15 | 36,256,723 |
Table 2: Number of arcs per instance.